Validate data quality of Phlogiston
Closed, ResolvedPublic
Actions

Assigned To

Authored By

	• JAufrecht
	Sep 17 2015, 9:19 PM

Description

Validate that the Phlogiston reports are accurate, especially the data used for forecasting.

Could do this by comparing to Phragile reports.
Could also create a known data set and then reporting on it. If we did this, then we could be sure Phlogiston was correct and use it to validate Phragile. - CUT and SPIN OFF

Related Objects

Mentioned Here: T114426: Validate data quality of Phlogiston via test data set

Event Timeline

• JAufrecht created this task.Sep 17 2015, 9:19 PM

• JAufrecht raised the priority of this task from to Medium.

• JAufrecht updated the task description. (Show Details)

• JAufrecht added a project: Team-Practices.

• JAufrecht subscribed.

Restricted Application added a subscriber: Aklapper. · View Herald TranscriptSep 17 2015, 9:19 PM

• Awjrichards moved this task from To Triage to General backlog on the Team-Practices board.Sep 17 2015, 9:26 PM

Notes from spot comparison of Phlogiston and Phragile spot:

Discrepencies:
Phragile's scope line doesn't start until mid-August. This is because Phragile does a nightly snapshot, rather than reconstruction, for scope, and Kevin associated Phragile with this project in mid-August.

Phragile burnup shows a spike today that is not in Phlogiston. This is because Phlogiston is running from a nightly dump.

Phragile scope line extends backwards and forwards to the end of the chart, but this does not represent data; this is just an extension of the edge values.

Phragile's burnup line is a historical reconstruction of point totals, but working from a list of tasks belonging to the project today. So if you remove a task from the project on Tuesday, on Wednesday Phragile's burnup will not show that task at any point, whereas Phlogiston's burnup will show that task up until Tuesday.